Linking problem in 7.10G

[arthur kogan]

Hi Bob,

I had a conversation with Alex Ling exactly about this issue couple of days ago
and we came up with exactly the same explanation. I have seen this problem
before (while I was working for Cognos support in Sydney), but in my case it was
linking a char*10 field to a char*8 field. If the value in the char*10 field was
9 or 10 characters long, you got a data conversion error.

I agree that this is a BAD programing technique, however in Alex's case (and
probably a lot of other people's) he is working on an old system and it would
take a big effort to make appropriate changes. The change requires not just the
database changes and recompilation. Think of all the temps and defines that need
to be changed. Also you may have a problem of space in screen and report layouts
if you have used the default pictures, as the short takes less space than long.
It could require a great deal of work to fix.

I therefore suggest that Cognos (rather than rejecting this "bug"), handle it in
a more user friendly way as follows. At compile time flag an appropriate
warning. At run time, do a pre-check on the value before trying to link, and in
case the value is to big, again flag a warning, not an error!!! That is, this
should be handled as "no match found" (which is the actual situation, since it
is impossible for a match to exist), but not an error stopping the program
running.

This way you are truly putting the user in control.

How about it?

Regards,

Arthur Kogan
Westpac Financial Services
Sydney, Australia

Deskin, Bob wrote:

> I just heard from Allison about the linking problem in 7.10G and I thought
> everyone would be interested. The bug # is 245681 and has to do with getting
> a data conversion error linking an InterBase long binary to an InterBase
> short binary. Turns out that the error existed at least as far back as
> 7.10E4 (and probably further back) but it was ignored rather than being
> reported. During the development of the QTP Tracer (which is Allison's baby
> and accomplishment - everyone say thank you to Allison), the code was
> changed to report the error so it could be traced. Now the clincher. The
> error only occurs when a value that is too large for the short binary is
> used. In other words, this is a valid error message. This bug is being
> rejected.
>
> Bob Deskin
> Senior Product Advisor  bob.deskin@cognos.com
> Cognos Inc.             (613) 738-1338 ext 4205 FAX: (613) 228-3149
> 3755 Riverside Drive P.O. Box 9707 Stn. T, Ottawa ON K1G 4K9 CANADA
> = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
> Subscribe: "subscribe powerh-l" in message body to majordomo@lists.swau.edu
> Unsubscribe: "unsubscribe powerh-l" in message to majordomo@lists.swau.edu
> powerh-l@lists.swau.edu is gatewayed one-way to bit.listserv.powerh-l
> This list is closed, thus to post to the list, you must be a subscriber.



= = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
Subscribe: "subscribe powerh-l" in message body to majordomo@lists.swau.edu
Unsubscribe: "unsubscribe powerh-l" in message to majordomo@lists.swau.edu
powerh-l@lists.swau.edu is gatewayed one-way to bit.listserv.powerh-l
This list is closed, thus to post to the list, you must be a subscriber.